In this fictionalized case, the protagonist is the director of talent and recruitment at Amazon and had been in charge of developing an artificial intelligence (AI) hiring tool that could shape the future of recruiting practices throughout the world. However, she begins to question the ethical implications of this tool. While the employees hired via this AI tool were performing exceptionally well, almost all of them were male, and she wonders how the tool she had helped create led to this result. There was no way to get around the fact that the hiring tool was gender biased, and the options are limited. The company could continue to use the tool as it currently existed, undoubtedly leading to a promotion for the protagonist. Or she could insist that Amazon invest more into the research and development of the AI tool, in hopes of creating an unbiased product. However, this option would be exorbitantly expensive in terms of both time and money and would reflect poorly on her efforts to produce this tool. A third option was for Amazon to scrap the AI hiring program altogether and return to traditional methods. The use of AI and automation was controversial, so perhaps showing the public that Amazon was putting people first would improve the company's image. The protagonist had to make a recommendation to senior management soon, and was uncertain what that should be. Excerpt UVA-E-0470 Feb. 2, 2021 Automated Hiring at Amazon It was 2:00 p.m., and you were sitting at your desk preparing for the meeting that could make or break your career. In your six short months at Amazon as director of talent and recruitment, you had been in charge of developing an artificial intelligence (AI) hiring tool that could shape the future of recruiting practices throughout the world. The tool, which had been piloted at Amazon Crystal City for the past quarter, had the ability to process thousands of resumes in minutes and flag the most qualified candidates. You had a meeting with Chuck Mudd, your boss and director of human resources at Amazon Crystal City, at 3:00 p.m. to evaluate the tool's performance and discuss next steps. As an African-American businesswoman and graduate from the Wharton School of Business, you knew the importance of hard work to succeed, especially in a male-dominated sector. After a long and successful career as a business technology consultant at McKinsey & Company, you had decided to shift career paths when you saw an opening for director of talent and recruitment at Amazon's second headquarters. Mudd had been especially impressed by your business-technology background and your ideas around automating Amazon's hiring process. Shortly after the start of your contract, Mudd gave you the green light to work alongside Arthur Mortimer, Amazon's AI director, to develop a tool to automate the rsum-review process. Mudd gave you free reign to determine the specifications of the bot, and proposed a one-quarter trial period for it. If it produced successful results, he promised a nice bonus and potential promotion. In developing the rsum-processing software, Mortimer fed the bot real resumes of tens of thousands of Amazon applicants, some of whom had been hired and some of whom had not. The bot learned the key qualities of the strongest rsums, such as the best universities, past position titles, and verbs associated with strong accomplishments. . . .
Read full abstract